| 00:39.13 | pra5ad | sean |
| 03:02.34 | *** join/#brlcad ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) | |
| 03:02.34 | *** topic/#brlcad is http://brlcad.org/ || BRL-CAD is now Free Software! || 7.6.4 is released || Developer meeting on Friday Nov 17! | |
| 03:19.17 | *** join/#brlcad Twingy (n=justin@pcp0011649600pcs.aberdn01.md.comcast.net) | |
| 03:57.22 | *** join/#brlcad fenn (n=4462aa53@bz.bzflag.bz) | |
| 05:28.36 | *** join/#brlcad DTRemenak|RDP (n=DTRemena@DHCP-170-143.caltech.edu) | |
| 05:40.18 | *** join/#brlcad autoeagle (n=9fe23f82@bz.bzflag.bz) | |
| 05:43.46 | autoeagle | hi,is any body here? |
| 05:44.27 | *** join/#brlcad autoeagle (n=9fe23f82@bz.bzflag.bz) | |
| 05:44.45 | brlcad | hello |
| 05:45.09 | brlcad | someone's always here, we're just not always quick to respond ;) |
| 05:47.33 | autoeagle | But I can't find any one or any information in the channel. |
| 05:49.20 | autoeagle | I have a question about the usage of the Brlcad. |
| 05:52.55 | autoeagle | If there are an existent function or software in the brlcad that could determine a point in or out of the model which build by brlcad? |
| 05:54.19 | autoeagle | If not, i want try to develop one for my project. |
| 05:55.36 | autoeagle | Beside, how to use the irc? |
| 05:57.49 | learner | determine a point in or out of the model? |
| 05:58.03 | learner | like shooting a ray and determining the entry and exit hit point? |
| 05:58.11 | learner | that's exactly what brl-cad is designed around ;) |
| 05:59.33 | brlcad | there are various net ettiquite rules for irc, the most important is probably to just be patient and ask well thought out questions if you have questions |
| 06:00.13 | brlcad | the conversation comes and goes so you pretty much stay in the channel while you do other things and check it from time to time |
| 06:01.23 | brlcad | as for your question about determining the in and out points of an intersection, it also matters whether you mean in the modeler or in code (both of which are possible |
| 06:02.06 | brlcad | in mged, you can run nirt or shoot_ray to fire a single ray (default is from the view center) and it will report some default hit information |
| 06:02.38 | brlcad | in code, rt_shootray() is where it all happens, and you register callbacks that get called on a hit/miss at which point you can report in/out points, etc |
| 06:03.04 | brlcad | e.g. look at src/rt/rtexample.c in the source distribution for a simple example that reports in and out hit points |
| 06:03.14 | autoeagle | In fact,it is not a very effective way to determine by shooting a ray and determining the entry and exit hit point. |
| 06:03.50 | brlcad | not effective for what purpose? |
| 06:04.23 | brlcad | and again, are you talking about code or user tools in brl-cad or just theoretically? |
| 06:05.35 | learner | most who use irc don't use that web interface, there are binary apps available for most all platforms (http://irchelp.org is a decent starting point) |
| 06:06.14 | autoeagle | In my project , the in or out test must be done more than billions times for a model. |
| 06:07.51 | autoeagle | In some case , such as the solid base on mesh, the ray test is effective for in and out test. |
| 06:07.57 | brlcad | billions isn't too bad, that's on par with the analysis simulations that are written to use brl-cad |
| 06:08.48 | brlcad | brl-cad csg ray-tracer can handle several million rays per second on a decent machine |
| 06:09.31 | brlcad | and if that's not fast enough, there is a triangle/mesh-based ray-tracer that can handle considerably more in brl-cad |
| 06:09.38 | autoeagle | And my model base on about 140 solid. |
| 06:09.50 | brlcad | that's a tiny model :) |
| 06:10.45 | brlcad | most of the models we're used to dealing with have thousands |
| 06:10.55 | brlcad | or tens of thousands |
| 06:11.03 | autoeagle | I think so. |
| 06:13.29 | brlcad | so it sounds then that you're interested in the programmatic interfaces in brl-cad? |
| 06:14.55 | autoeagle | but my project include 1M particles. each one need the in or out test per turn.And the whole process includes more than 100,000 turn, |
| 06:16.32 | brlcad | that should be pretty straight-forward to implement |
| 06:16.51 | brlcad | though if it's a particle system you're implementing, you probably don't need full-blown ray-tracing |
| 06:17.19 | brlcad | you just need simple locale-specific collision detection and velocity updates |
| 06:17.35 | brlcad | that would be a cool feature to add to brl-cad |
| 06:17.53 | autoeagle | Yes |
| 06:17.55 | brlcad | there's minimal particle system and time-based update support |
| 06:19.03 | brlcad | that said, you certainly 'could' implement the particle system using the ray-tracer pretty darn easily |
| 06:20.15 | autoeagle | theoretically, a simple locale-specific collision could save 50-90% calculate time vs ray test. |
| 06:20.28 | Maloeran | Adding code to break out of adrt's tracing loop after travelling a certain distance wouldn't be much trouble, for local collision tests |
| 06:20.36 | brlcad | you'd shoot a ray for every particle, every timestep to determine if/where they intersect -- it would take about 1s per timestep on a desktop |
| 06:21.08 | brlcad | that'd be about a 26 hour computation on a desktop |
| 06:21.16 | brlcad | for 100k timesteps |
| 06:21.38 | autoeagle | really? |
| 06:21.39 | brlcad | you can undoubtedly improve upon that given the simple nature of the sim |
| 06:23.32 | brlcad | the main killer time sink by ray-tracing is probably going to be prep time since the model changes every time step, i'd bet that'd account for most of the time without some customization |
| 06:25.28 | autoeagle | Yes,In my project , the rebuild of the list will cost about 90% calculations. |
| 06:26.26 | brlcad | what are your time constraints? what's the goal? an animation of the simulation? a (real-time?) simulation framework? |
| 06:27.45 | autoeagle | not real time. |
| 06:28.09 | brlcad | no sense worrying about whether it's going to take 26 hours or 40 hours if real-time is a requirement; vice-versa if real-time isn't a requirement and you're just rendering off-line, no bid deal if it simulates and renders for a day |
| 06:30.36 | autoeagle | but the now effection is too bad.So i want to optimize it with brlcad. |
| 06:31.25 | autoeagle | the first step the change the modeling system to mged |
| 06:31.45 | brlcad | if you represent your particles with spheres or triangles, you should get optimal raytrace performance using rt or adrt respectively |
| 06:31.50 | autoeagle | the first step is to change the modeling system to mged |
| 06:32.06 | brlcad | please do :) |
| 06:36.59 | brlcad | in the meanwhile, presuming you're going to hang around (which I hope you do) -- I'd suggest http://www.irssi.org/ or http://www.xchat.org/ for an irc client |
| 06:37.28 | autoeagle | thank you:) |
| 06:40.22 | autoeagle | which Networks should I choice in Xchat. |
| 06:42.06 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/NEWS: it's 7.6.5, not 7.6.6 quite yet.. |
| 06:42.08 | brlcad | irc.freenode.net |
| 06:42.23 | brlcad | #brlcad |
| 06:43.28 | brlcad | (port 6667) |
| 06:43.34 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/TODO: fix perf.sh pathing issue |
| 06:43.40 | brlcad | fg |
| 06:45.22 | CIA-12 | BRL-CAD: 03brlcad * 10brlcad/TODO: test pic/nonpic build problems with librtms/libmultispectral |
| 06:48.42 | brlcad | welp, time for me to sleep a lil bit I think |
| 06:48.55 | brlcad | nice talking to you autoeagle, see you around more hopefully |
| 06:50.36 | autoeagle | thank you. |
| 06:51.21 | autoeagle | bye |
| 06:52.32 | autoeagle | blushing for my poor english. |
| 07:05.54 | *** join/#brlcad cad954 (n=9fe23f82@bz.bzflag.bz) | |
| 07:06.12 | *** join/#brlcad autoeagle (n=9fe23f82@bz.bzflag.bz) | |
| 07:07.16 | *** join/#brlcad cad694 (n=9fe23f82@bz.bzflag.bz) | |
| 07:12.45 | *** join/#brlcad cad575 (n=9fe23f82@bz.bzflag.bz) | |
| 08:13.56 | *** join/#brlcad DTRemenak|RDP (n=DTRemena@DHCP-170-143.caltech.edu) | |
| 09:24.35 | *** join/#brlcad learner (n=brlcad@pcp0011649376pcs.aberdn01.md.comcast.net) | |
| 09:24.35 | *** mode/#brlcad [+o learner] by ChanServ | |
| 12:21.45 | *** join/#brlcad brlcad_ (n=brlcad@bz.bzflag.bz) | |
| 12:30.00 | *** join/#brlcad archivist (n=archivis@host217-35-76-52.in-addr.btopenworld.com) | |
| 13:51.07 | *** mode/#brlcad [+o brlcad] by ChanServ | |
| 15:49.37 | *** join/#brlcad Maloeran (n=alexis@modemcable084.164-81-70.mc.videotron.ca) | |
| 18:58.58 | *** join/#brlcad cad009 (n=0c0b8913@bz.bzflag.bz) | |
| 19:03.04 | *** join/#brlcad DTRemenak (n=DTRemena@DHCP-170-143.caltech.edu) | |
| 19:10.04 | *** join/#brlcad DTRemenak (n=DTRemena@DHCP-170-143.caltech.edu) | |
| 20:59.36 | *** join/#brlcad raz (n=raz@pool-138-88-142-43.esr.east.verizon.net) | |
| 21:01.47 | raz | twingy: we are running the isst apps but not seeing anything rendered. we followed the usage steps in the readme. Any input? |
| 21:12.06 | ``Erik | is it printing the framerate in the console? |
| 21:27.49 | raz | the splash screen shows up and the caption reads: ADRT_ISST_Observer Loading... |
| 21:37.43 | ``Erik | hrm, lemme go kick justin and make him look at irc o.O |
| 21:39.47 | raz | thanks |